Getting Started with PromptQuery: AI Data Analysis and Database Management with MCP

PromptQuery is an MCP-native desktop database client for AI-powered data analysis and database management. It lets you connect SQL and NoSQL databases, publish selected connections through a local MCP server, and let AI agents help query data, analyze results, generate reports, browse schemas, and manage database workflows safely.

This guide walks through the current PromptQuery workflow: connect a database, explore it manually, enable MCP access, ask an AI agent for analysis, and keep risky operations under human control.

What You Can Do with PromptQuery

PromptQuery combines a traditional database client with AI-agent workflows.

You can use it to:

Connect to SQL and NoSQL databases
Browse schemas, tables, collections, and rows
Run SQL queries and inspect results
Import and export CSV files
Generate ER diagrams from live schemas
Edit rows and batch-commit changes
Publish a database connection as a local MCP server
Let AI agents analyze data and generate reports
Review risky operations before they touch your database

Instead of using one tool for database management and another for AI analysis, PromptQuery brings both workflows together.

Who This Guide Is For

This guide is useful if you want to:

Ask Claude, Cursor, Codex, Windsurf, GitHub Copilot, Antigravity, OpenCode, or another MCP-compatible agent to work with your database
Analyze data using natural language
Generate SQL without manually writing every query
Create reports or charts from live data
Browse and manage databases from a desktop client
Keep AI database access safe and reviewable

You do not need to be a database expert to get started, but you should have access to a database you are allowed to query.

Step 1: Install PromptQuery

Download PromptQuery for your operating system:

macOS
Windows
Linux

After installation, open the desktop app and sign in or create an account if required by your plan.

Download PromptQuery

Step 2: Connect Your Database

The first step is adding a database connection.

PromptQuery supports many SQL and NoSQL databases, including:

PostgreSQL
MySQL
SQL Server
Oracle
SQLite
MariaDB
Supabase PostgreSQL
MongoDB and MongoDB Atlas
Redis
DynamoDB
Firebase

To connect a database:

Open PromptQuery.
Go to the database connection area.
Choose your database type.
Enter the required connection details.
Test the connection.
Save the connection.

Once connected, PromptQuery can read your schema, show tables or collections, inspect columns and data types, and prepare useful context for AI-assisted workflows.

Step 3: Explore Your Database Manually

Before involving an AI agent, it is often helpful to explore the database directly in PromptQuery.

You can:

Browse schemas and tables
Preview rows
Inspect columns and data types
Run SQL queries
View query results
Export data when needed
Generate ER diagrams to understand relationships

This gives you confidence that the connection is correct and helps you understand what data is available.

Step 4: Enable MCP for AI Agents

PromptQuery includes a built-in MCP server.

MCP lets AI agents use structured tools and context from external applications. In PromptQuery, this means an AI agent can ask for database schema information, generate queries, run safe read operations, and help analyze results through a controlled local interface.

A typical MCP workflow looks like this:

Connect your database in PromptQuery.
Enable or publish the connection as an MCP server.
Connect your MCP-compatible AI agent.
Ask the agent a database question.
Review queries, results, reports, and any risky operations in PromptQuery.

This is safer than pasting database credentials directly into an AI tool because PromptQuery remains the control layer between the agent and your database.

Step 5: Connect an AI Agent

PromptQuery is designed to work with MCP-compatible AI tools and agents, including:

Claude
Cursor
Codex
Windsurf
GitHub Copilot
Antigravity
OpenCode
Other MCP-compatible agents

After connecting your agent to PromptQuery's MCP server, the agent can use database context without you manually copying schemas, table definitions, or query results into chat.

Step 6: Ask Your First Data Question

Start with a simple question that only requires read access.

For example:

Show me the top 10 customers by revenue this month.

Or:

Analyze signups over the last 30 days and summarize the trend.

Or:

Find tables related to billing, payments, or subscriptions and explain how they connect.

The agent can use PromptQuery to inspect the schema, identify relevant tables, generate SQL, run safe read queries, and summarize the result.

Step 7: Review Generated SQL and Results

Even with AI assistance, you should review database operations.

When an agent generates SQL, check:

Does it query the correct tables?
Are the selected columns relevant?
Are filters and date ranges correct?
Are joins using the right relationships?
Does the result answer the original question?

PromptQuery is designed for human-in-the-loop workflows. AI can help move faster, but you stay in control of what runs and what changes.

Step 8: Generate a Report

PromptQuery can help turn database questions into shareable analysis.

For example, ask your agent:

Create a report showing monthly revenue by plan, explain the biggest changes, and include a chart.

The agent can use PromptQuery's database context to generate the query, analyze returned data, and produce a report based on your real database.

This is useful for:

Product analytics
Revenue analysis
Customer operations
Internal reporting
Debugging product behavior
Sharing findings with teammates

Step 9: Use Database Management Features

PromptQuery is also a full database management tool.

You can continue working manually after the AI-assisted analysis:

Browse table data
Edit rows
Import CSV files
Export query results
Generate ER diagrams
Review changes before committing
Batch-commit inserts, updates, and deletes

This means you do not need to switch between a database client, an AI chat tool, and a reporting tool for every workflow.

Step 10: Keep Risky Operations Safe

PromptQuery is safe by default.

AI agents can inspect and query databases, but risky operations such as DELETE, DROP, and UPDATE require explicit approval.

Recommended safety practices:

Start with read-only access for every new connection.
Review generated SQL before execution.
Use least-privilege database users when possible.
Avoid connecting production databases with broad write permissions.
Require approval for destructive or schema-changing operations.
Test workflows on development or staging databases first.

The goal is not to block AI agents from being useful. The goal is to make sure useful agent workflows happen inside clear safety boundaries.

Example Workflow: Revenue Analysis

Here is a practical first workflow to try.

Ask your AI agent:

Analyze revenue for the last 90 days, group it by month and plan, identify the biggest changes, and create a short report.

With PromptQuery, the agent can:

Inspect available schemas and tables.
Find billing, subscription, invoice, or payment data.
Generate the SQL query.
Run safe read-only analysis.
Summarize the result.
Create a chart or report.
Help refine the analysis with follow-up questions.

You can then review the SQL, inspect results in PromptQuery, export data if needed, and share the report.

Tips for Better AI Database Results

To get better results from AI agents, write prompts with enough context.

Good prompts include:

The business question you want answered
The time range
The metric you care about
The grouping or breakdown you want
Any known table or column names
The desired output format

For example, this is better:

Analyze paid subscription revenue for the last 6 months, grouped by plan and month. Exclude refunded payments. Return SQL, a short summary, and a chart-friendly result table.

Than this:

Show revenue.

The more specific your request, the easier it is for the agent to generate useful queries and analysis.

What to Try Next

After your first query or report, try these workflows:

Generate an ER diagram for your database.
Ask an agent to explain a complex table.
Export query results to CSV.
Import a CSV into a test table.
Ask for a report about product usage trends.
Use PromptQuery with a staging database before production.
Save useful queries for future analysis.

Conclusion

PromptQuery helps you use AI agents with databases safely.

You get a desktop database client for daily database management, plus an MCP-native bridge that lets AI agents query data, analyze results, generate reports, and understand schemas without bypassing human control.

Start with a read-only workflow, ask a focused data question, review the generated SQL, and build from there.

Download PromptQuery